Ansible Roles: Mesosphere DC/OS

This is a collection of Ansible Roles to manage a DC/OS cluster on RedHat/CentOS Linux.

Requirements

To effectively use these roles, your nodes should follow the setup recommended by Mesosphere. Depending on your configuration, you need to deploy to:

• One or more master nodes ('masters')
• One bootstrap node ('bootstraps')
• Zero or more agent nodes for public services ('agents_public')
• One or more agent nodes for private services ('agents_private')

Example Inventory File

Here’s an example of how your inventory file might look:

[bootstraps]
bootstrap1-dcos112s.example.com

[masters]
master1-dcos112s.example.com
master2-dcos112s.example.com
master3-dcos112s.example.com

[agents_private]
agent1-dcos112s.example.com
remoteagent1-dcos112s.example.com

[agents_public]
publicagent1-dcos112s.example.com

[agents:children]
agents_private
agents_public

[common:children]
bootstraps
masters
agents
agents_public

Role Variables

The Mesosphere DC/OS Ansible roles use two types of variables:

1. Per-node type variables (group_vars)
2. A multi-level dictionary called dcos, available to all nodes

Per Group Variables

[bootstraps:vars]
node_type=bootstrap

[masters:vars]
node_type=master
dcos_legacy_node_type_name=master

[agents_private:vars]
node_type=agent
dcos_legacy_node_type_name=slave

[agents_public:vars]
node_type=agent_public
dcos_legacy_node_type_name=slave_public

Global Variables

dcos:
  download: "https://downloads.dcos.io/dcos/stable/1.13.4/dcos_generate_config.sh"
  download_checksum: "sha256:a3d295de33ad55b10f5dc66c9594d9175a40f5aaec7734d664493968a9f751fd"
  version: "1.13.4"
  enterprise_dcos: false
  selinux_mode: enforcing

  config:
    cluster_name: "examplecluster"
    security: strict
    bootstrap_url: http://int-bootstrap1-examplecluster.example.com:8080
    exhibitor_storage_backend: static
    master_discovery: static
    master_list:
      - 172.31.42.1

Cluster-wide Variables

Name	Required?	Description
download	REQUIRED	(https) URL to download the Mesosphere DC/OS installation script.
download_checksum	no	Checksum to verify the download. Starts with the method used (e.g., "sha256:").
version	REQUIRED	The version string of the installer (found at download). You can get this by running dcos_generate_config.sh --version.
version_to_upgrade_from	for upgrades	The version string of Mesosphere DC/OS expected to upgrade FROM. The upgrade script will be made on the bootstrap machine, and each cluster node will download the correct upgrade for its current DC/OS version.
image_commit	no	Used for consistent version/config upgrades, particularly for non-released versions (e.g., 1.12-dev). This takes precedence over version.
enterprise_dcos	REQUIRED	Specifies if the installer (found at download) installs an 'open' or 'enterprise' version. This is important as enterprise components have extra post-upgrade checks.
selinux_mode	REQUIRED	Indicates the SELinux mode for cluster nodes. Mesosphere DC/OS supports enforcing mode from 1.12 onward; older versions require permissive.
config	yes	YAML structure for a valid Mesosphere DC/OS config.yml.

DC/OS config.yml Parameters

Refer to the official Mesosphere DC/OS configuration documentation for a complete list of parameters. Key parameters include:

• bootstrap_url: Should point to http://your bootstrap node:8080. This will be used internally and conveniently overwritten for the installer/upgrader to direct to a version-specific sub-directory.
• ip_detect_contents: User-supplied IP detection script. Overrides built-in environment detection and common AWS/on-premise scripts. DC/OS IP detection reference
• ip_detect_public_contents: User-supplied public IP detection script. Overrides built-in detection.
• fault_domain_detect_contents: User-supplied fault domain detection script. Also overrides built-in detection.

Ansible Dictionary Merge Behavior Note

Due to the nested structure of the dcos configuration, you may need to set Ansible to 'merge' instead of 'replace' when combining configurations from different sources.

Example

# ansible.cfg
hash_behaviour = merge

Safeguard for Interactive Use: dcos_cluster_name_confirmed

When running these roles interactively, the DCOS.bootstrap role will require confirmation of the cluster name to run against. This is a failsafe against unintentional upgrades or configuration changes. In automated plays, set a variable to skip this, for example:

ansible-playbook -e 'dcos_cluster_name_confirmed=True' dcos.yml

Example Playbook

Mesosphere DC/OS is a complex system, comprising multiple nodes to form a full multi-node cluster. There are some considerations when creating a playbook using the provided roles:

1. The order in which groups run their roles (e.g., bootstrap first, then masters, then agents).
2. Control over concurrency for upgrades (e.g., use serial: 1 for master nodes).

You can use the provided dcos.yml playbook to install and upgrade Mesosphere DC/OS.

Tested Operating Systems and DC/OS Versions

• CentOS 7, RHEL 7
• DC/OS 1.12, both open source and enterprise versions

License

Apache 2.0

Author Information

This role was created by the SRE team at Mesosphere and others in 2018, based on various internal tools and Ansible roles developed over the years.